5:15 PM - 6:30 PM
[SSS09-P04] Migration of fluid during oil recovery revealed by Reverse Time Imaging
The goal of microseismic source location method is to determine the spatial location and time of origin of a seismic event (Thurber 2011), which has significant value in the study of active faults, earth's internal structure, and shale gas production. Traditional seismic source location method mainly depends on the first arrival time, which requires picking the first arrival time accurately. Such seismic event location methods have limitations, because it is difficult to pick up the first arrival time in the records of the microseismic events that has low signal-to-noise (S/N) ratio. Reverse Time Imaging (RTI) method is an alternative event location method that is based on reconstructing wavefront and constructive stacking of seismic waveforms. The RTI method does not need first-arrival time and is suitable for data with low S/N (Gajewski and Tessmer 2005).
After its invention, the RTI method was continuously developed and broadly used. Steiner et al. (2008) used low-frequency information to identify reservoirs locations and provides a possible method for reservoir localization. Zou et al. (2014) successfully identified a doublet of earthquakes in the TGR region. Zhu (2014) discussed the RTI using the acoustic wave equation in attenuation media and proposed a method to overcome the influence of attenuation on the result of event location. Ge et al. (2019) combined the travel time location and RIT methods, and proposed the reverse travel time imaging (RTTI) method. Zhu (2019) developed a hybrid multiplicative time-reversal imaging (HyM-TRI) algorithm, for automatically tracking the spatiotemporal distribution of microseismic events. For its applicability in low SNR situation, the RTI has potential to image the noise generated by the movement of fluid in petroleum reservoir.
In this paper, RTI is applied to study the dynamic evolution of subsurface fluid during the steam injection for petroleum production. The real data before and after steam injection are processed respectively, and then compared to show the changes beneath the oil field. Since the noise surrounding the petroleum reservoir generated by the injection process usually has a relatively stable energy in a period of several minutes, the data of the two parts are superimposed in the time domain to suppress the influence of the random noise. Two images that are before and during the injection are normalized into the same background, and then subtracted with each other to show the difference that reflects the microseismic information generated by fracturing in the steam injection process.
After its invention, the RTI method was continuously developed and broadly used. Steiner et al. (2008) used low-frequency information to identify reservoirs locations and provides a possible method for reservoir localization. Zou et al. (2014) successfully identified a doublet of earthquakes in the TGR region. Zhu (2014) discussed the RTI using the acoustic wave equation in attenuation media and proposed a method to overcome the influence of attenuation on the result of event location. Ge et al. (2019) combined the travel time location and RIT methods, and proposed the reverse travel time imaging (RTTI) method. Zhu (2019) developed a hybrid multiplicative time-reversal imaging (HyM-TRI) algorithm, for automatically tracking the spatiotemporal distribution of microseismic events. For its applicability in low SNR situation, the RTI has potential to image the noise generated by the movement of fluid in petroleum reservoir.
In this paper, RTI is applied to study the dynamic evolution of subsurface fluid during the steam injection for petroleum production. The real data before and after steam injection are processed respectively, and then compared to show the changes beneath the oil field. Since the noise surrounding the petroleum reservoir generated by the injection process usually has a relatively stable energy in a period of several minutes, the data of the two parts are superimposed in the time domain to suppress the influence of the random noise. Two images that are before and during the injection are normalized into the same background, and then subtracted with each other to show the difference that reflects the microseismic information generated by fracturing in the steam injection process.